The Poor and Their Money An essay about financial services for poor people pdf

i
The Poor and Their Money
An essay about financial services for poor people
Stuart Rutherford
Institute for Development Policy and Management
University of Manchester
January 1999
The Department for International Development will be publishing this work in
New Delhi during 1999. For further information contact Sukhwinder Arora at the
Department for International Development, New Delhi, India.
ii
PREFACE
Over the last 15 years initiatives to provide financial services to poor people (the ‘microfinance
industry’) have come on by leaps and bounds in terms of size and reputation. Despite this, the
industry is still only in its adolescence and our understanding of why and how poor and very poor
people use microfinancial services ( and why many choose not to use the services that are available)
remains partial at best. This essay takes the reader on a ‘voyage of discovery’ that seeks to both
deepen her/his understanding and encourage her/him to apply that knowledge to the practice of
microfinance.
The voyage that Stuart Rutherford offers is a unique one based upon years of careful and detailed
personal research. It does not take a deductive approach that develops a theoretical model of the
financial behaviour of poor people. Nor does it follow the ‘case study plus best practice’ approach that
has been favoured by many practitioners when they write of microfinance. Instead, it adopts an
inductive approach - based on thousands of conversations and meetings with poor people discussing
what financial services they use and need - backed up by the personal experience of running an
experimental microfinance institution (SafeSave).
This is an innovative methodology and one that courts considerable risks, not least that it might
produce many interesting insights but no clear analysis. However, by tempering reflection with action
research these risks are avoided and the essay provides a wealth of thought provoking empirical
observations woven into an insightful analysis of how microfinance institutions (MFIs) can operate
more effectively and how they might develop products that more adequately meet the needs of the

market niche that they seek to fill. These products must meet the many differing needs of poor people
- for savings, insurance, productive investment, and housing, education, health and consumption
loans - and should be designed so that they are convenient, secure, flexible and low cost. Such
services can help poor and very poor people cope with the vulnerability of their livelihoods and
support their personal efforts to achieve the economic and social goals that they set for themselves.
The argument for providing microfinancial services in this essay is not the conventional microcredit
argument - that small loans to poor microentrepreneurs puts them on a conveyer belt that takes them
over the ‘poverty line’. Rather, it is altogether more subtle and more grounded in the reality of being
poor - high quality microfinancial services can help the poor to help themselves to overcome their
problems and to seize the opportunities that they identify. While the author may be correct in saying
that this essay is not an ‘academic paper’ in a conventional sense, it is the work of a scholar-
practitioner (or practitioner-scholar) who knows the academic debates very well, who has a detailed
knowledge of contemporary MFIs, who knows about the wealth of informal microfinancial services that
the poor use and who runs an MFI on a day-to-day basis. It challenges the assumptions that
underpin much of the academic analysis of microfinance within the field of development studies and
moves on to provide guidance (but not blueprints) about how to improve practice.
Those who know Stuart’s ideas already will be pleased to see them brought together in this working
paper. Those who are not yet acquainted with his ideas are in for a real treat - sit down in comfort,
take the phone off the hook, and read on.
David Hulme
Manchester
February 1999
iii
The Poor and their Money
An essay about financial services for poor people
Introduction
This essay is about how poor people in developing countries manage their money. It describes how
they handle their savings, from keeping bank notes under the floorboards to running sophisticated
savings and loan clubs. It illustrates the variety of moneylenders and deposit collectors who serve the
poor, including the new breed of ‘microfinance institutions’ (MFIs) - semi-formal or formal banks that

specialise in working with poor clients.
The essay illustrates the principles that underlie these phenomena, and by doing this I hope it will
stimulate those who would like to improve the quality of financial services that are offered to the
poor.
In short, the essay is about how a better understanding of financial services for the poor can lead to
better provision of such services.
The audience I have in mind is composed mainly of those who provide or promote financial services
for the poor, and their backers. I am thinking of men and women who work for MFIs and non-
government organisations (NGOs), aid donors, and banks and co-operatives who want to reach the
poor. I hope, too, that some members of the general public will find the essay interesting and
readable.
The essay therefore aims at clarity. I try to avoid jargon. Academic machinery such as footnotes and
references is used as sparingly as possible, though there is a bibliography which has been
annotated to help practitioners. Most of the cases that I use to illustrate my points are ones that I
have personally investigated during more than twenty years of research and practice in the subject
on three continents (though there is a strong bias towards Asia, where I have lived and worked for
fourteen years).
‘The Poor and Their Money’ is not an academic paper. I hope some academics will read the essay,
because they too influence the growing ‘microfinance industry’, but they should not expect it to
conform to academic standards of presentation and argument. Many of the statements I make are
grounded on my long-standing interest and experience in the field, above all on my conversations
with poor people about how they actually use financial services. I have not made any assumptions
that are not based on this kind of experience. But in the interest of brevity and readability I have not
quoted chapter-and-verse in support of all my arguments, as would be required in a formal
academic paper. I invite academics to get in touch with me (on Error! Reference source not found.Error! Reference source not found.)
if they would like more references, or if they would like to challenge or amplify what I have written.
Nor is the essay intended as a ‘manual’. I do not provide step-by-step guidance in how to set up an
MFI. Although I describe my own work – the MFI called SafeSave that features at the end of chapter
one – I don’t for one minute think that SafeSave is the last word in financial services for the poor.
SafeSave is included in the essay to illustrate some important issues, and not as a recommended

‘recipe’. Indeed, by the time you read this, SafeSave should have moved on to new and – we hope
- better products. SafeSave happens to be my ‘action research’ project, and I would encourage
others to set up research vehicles of their own.
The ‘microfinance industry’ is in its adolescence. There have been encouraging breakthroughs in the
last two or three decades – as Chapter Three shows. But the potential for growth and improvement is
iv
huge. There are still millions of poor people to reach, and hundreds of new ways of reaching them
waiting to be discovered and developed. I hope this essay will accelerate this voyage of discovery.
Acknowledgements
I was persuaded to write this essay by Sukhwinder Singh Arora, of the UK government's 'Department
for International Development' (DFID) in Delhi. Much of the material I use was uncovered in
Sukhwinder's company in cities around India in the course of work for DFID. To him I owe a double
vote of thanks. Graham Wright, now working for DFID and UNDP in East Africa, is another co-
researcher who encouraged me, and who tramped through villages with me in Bangladesh and The
Philippines. Another who was closely involved in researching material for the essay is my assistant S K
Sinha. Help and encouragement has come from many sources. They include the many
organisations with whom I have worked. Although there are too many to list, I would like to pick out
ASA, ActionAid (especially in Bangladesh and Vietnam), BURO Tangail, CARE International (in several
countries), DFID, and PLAN International, as well as my own MFI SafeSave and my own academic
institution, the Institute for Development Policy and Management (IDPM) at the University of
Manchester. Those who have helped through discussion or through reading drafts of this essay (or
parts of it) include Edward Abbey, Dale Adams, Thierry Van Bastelaer, Gregory C Chen, Robert
Christie, Hege Gulli, Robert Hickson, David Hulme, Feisal Hussain, Sanae Ito, Susan Johnson, Vijay
Mahajan, Mahini Malhotra, Imran Matin, Jonathan Morduch, Rich Rosenberg, Hans Seibel, William
Steel, and Astrid Ursem. I have benefited from all of them, but while I am willing to share with them
the credit for any virtues the essay may have, I jealously guard my sole ownership of its faults.
Thousands of users and would-be users of financial services for the poor around the world have given
their time to teach me how and why the existence and quality of financial services is important to
them. Since it is hard to list or to thank them, I acknowledge my debt by dedicating this essay to
them.

Stuart Rutherford
Dhaka, Bangladesh, 1998
v
AbstractAbstract
Poor people can save and want to save, and when they do not save it is because of lack of
opportunity rather than lack of capacity. During their lives there are many occasions when they need
sums of cash greater than they have to hand, and the only reliable way of getting hold of such sums
is by finding some way to build them from their savings. They need these lump sums to meet life-
cycle needs, to cope with emergencies, and to grasp opportunities to acquire assets or develop
businesses. The job of financial services for the poor, then, is to provide them with mechanisms to
turn savings into lump sums for a wide variety of uses (and not just to run microenterprises). Good
financial services for the poor are those that do this job in the safest, most convenient, most flexible
and most affordable way.
The poor seek to turn their savings into lump sums by finding reliable deposit takers, by seeking
advances against future savings (loans), or by setting up devices like savings clubs and ROSCAs. A
study of these traditional methods reveals the importance of the frequency and regularity of
deposits, of the time-scale over which the deposit/lump-sum swap is made, and of the relative
merits of systems that offer just one kind of swap as against those that offer multiple swap types. It
also shows how interest rates have been used to manage the risks faced by savings club members.
Some, but not all, of these lessons have been learned by the two new sets of players that have
emerged recently to form the new 'microfinance industry'. There are 'promoters' - organisations that
seek to help the poor set up financial services devices owned by themselves or their communities -
and 'providers' - new financial intermediaries which sell financial products to the poor. Providers, it is
found, are better able to reach large numbers of poor people with innovative products that build on
the experience of the informal sector. To develop good financial services for the poor we need
products that suit the poors’ capacity to save and their needs for lump sums, and product delivery
systems that are convenient for the poor. The essay ends by discussing how the process of
establishing such products and institutions can be accelerated.
The essay is not an academic paper. It is aimed at microfinance practitioners and their backers,
and is intended to stimulate them to invent and test financial products for the poor and to develop

suitable institutions to deliver the products.
vi
Contents
Introduction
Acknowledgements
Abstract
Chapter One:Chapter One:
The Poor and Their MoneyThe Poor and Their Money
Why the poor need financial services: some examples of the services they get and of
how they use them
Chapter Two:
Doing it yourself: ROSCAs and ASCAs
How the poor run savings clubs
Chapter Three:
Using the informal sector: managers and providers
About the services that the poor can buy in the informal market
Chapter Four:
New ways to manage money: promoters and providers
About the new organisations that have sprung up recently to help the poor get better
financial services
Chapter Five:
Reprise: better financial services for the poor
Some concluding remarks
Further reading
An annotated list of books and articles for further reading
Chapter One: The Poor and Their Money
Three facts and a conclusion. Fact one: poor people can and do save, even if the
amounts are often small and irregular. Fact two: poor people need usefully large lump
sums of money from time to time, for many different purposes. Fact three: for most poor
people, those ‘usefully large lump sums’ have to be built, somehow or other, out of their

savings - because there is no other reliable way to get hold of them. Conclusion: financial
services for poor people are largely a matter of mechanisms that allow them to convert a
series of savings into usefully large lump sums.
vii
1 The poor as savers
The poor want to save, and do save… but it’s not easy
A popular and useful definition of a poor person is someone who doesn't have much money.
Among academics, and in the aid industry, this definition has gone out of fashion. But it suits our
present purposes well, so we shall stick to it. In this essay, when I talk about ‘the poor’, I mean people
who, compared to their fellow citizens, don't have much money.
If you don't have much money it is especially important that you manage well what money you
have. Poor people are at a disadvantage here, because the banks and insurance companies and
other financial institutions that serve the better-off rarely cater to the poor. Nevertheless, poor people
do seek and find a variety of ways of better managing their money, as examples in this essay will
show. The essay argues that we can learn a lot from the more successful money-managing efforts of
the poor, and use that learning to design new and better ways of bringing banking services to the
slums and villages of the developing world.
Choosing to save…
Managing money well begins with hanging on to what you have. This means avoiding unnecessary
expenditure and then finding a safe place to store whatever money is left over. Making that choice -
the choice to save rather than to consume - is the foundation of money management.
…but finding it hard to do so
Poor people run into problems with money management at this very first hurdle. If you live in an
urban slum or in straw hut in a village, finding a safe place to store savings is not easy. Bank notes
tucked into rafters, buried in the earth, rolled inside hollowed-out bamboo, or thrust into clay piggy
banks, can be lost or stolen or blown away or may just rot. Certainly their value will decline, because
of inflation. But the physical risks are the least of the problem. Much tougher is keeping the cash safe
from the many claims on it - claims by relatives who have fallen on hard times, by importunate
neighbours, by hungry or sick children or alcoholic husbands, and by landlords, creditors and
beggars. Finally, even when you do a have a little cash left over at the day’s end, if you don't have

somewhere safe to put it you’ll most probably spend it in some trivial way or other. I have lost count
of the number of women who have told me how hard it is to save at home, and how much they
would value a safe, simple way to save.
Nevertheless, the poor can save, do save, and want to save money. Only those so poor that they
have left the cash economy altogether - the elderly disabled, for example, who live by begging
food from neighbours - cannot save money. This essay is not about them.
But can the poor really save?
The fact that the poor want to save and have some capacity to save is not self-evident. If you don't
know much about how the poor actually organise their lives you may assume that the poor ‘are too
poor to save’. The poor spend all their income and still don't get enough to eat, so how can they
save? The poor may need loans, but the last thing they need, you may think, is a savings service.
Ins and outs
By the time you have finished this essay you should see that this is a misconception. But for the time
being, notice that people (and not just the poor) may save money as it goes out (keeping a few
coins back from the housekeeping money) as well as when it comes in (deducting savings at source
from your wage or other income). Even the poorest have to spend money to buy basic items like
food and clothing, and each time they do so there is the opportunity to save something, however
tiny. Many poor housewives try to save in this way, even if their working husbands fail to save anything
from their income.
That the poor do succeed in saving something is shown by their habit of lending each other small
amounts of money (as well as small amounts of rice or kerosene or salt). In this ‘reciprocal lending’ I
lend you a few cents today on the understanding that you’ll do the same for me at some other time.
The practice is so common that such loans make up the bulk of financial transactions that poor
people get involved in, even if the amounts involved add up to only a small proportion of the total
value in circulation in financial services for the poor. The practice depends entirely on the poor’s
capacity and willingness to save.
viii
This essay is about saving money. People save in other ways, of course, and we shall take that into
account, briefly, at the end of this chapter. But for the time being I want to pursue my basic
message in the simplest way, and that means concentrating on money savings. The poor, we have

claimed, can and do save. But why do they do so?
2 The poor as big spenders
The poor need, surprisingly often, to spend large sums of money
You may not yet be fully convinced that the poor can and do (and want to) save. So we shall move
on to the spending needs of the poor, which are less controversial.
The need to spend
Just because you’re poor doesn't mean that all your expenditure will be in small sums. Much of it
may be - you may buy only a little food or clothing at a time. But from time to time you need to
spend large sums. We can list these times in three main categories, ‘life-cycle’ events, emergency
needs, and investment opportunities.
Life-cycle needs
In Bangladesh and India, the dowry system makes marrying daughters an expensive business. In
parts of Africa, burying deceased parents can be very costly. These are just two examples of ‘life-
cycle’ events for which the poor need to amass large lump sums. Other such events include
childbirth, education, home-building, widowhood and old-age generally, and the desire to
bequeath a lump sum to heirs. Then there are the recurrent festivals like Eid, Christmas, or Diwali. In
each case the poor need to be able to get their hands on sums of money which are much bigger
than the amounts of cash which are normally found in the household. Many of these needs can be
anticipated, even if their exact date is unknown. The awareness that such outlays are looming on the
horizon is a source of great anxiety for many poor people.
Emergencies
Emergencies that create a sudden and unanticipated need for a large sum of money come in two
forms - personal and impersonal. Personal emergencies include sickness or injury, the death of a
bread-winner or the loss of employment, and theft or harassment. Impersonal ones include events
such as war, floods, fires and cyclones, and - for slum dwellers - the bulldozing of their homes by the
authorities. Again, you will be able to think of other examples. Each creates a sudden need for more
cash than can normally be found at home. Finding a way to insure themselves against such troubles
would help millions of poor people.
Opportunities
As well as needs for spending large sums of cash, there are opportunities to do so. There may be

opportunities to invest in an existing or new business, or to buy land or other productive assets. The
lives of some poor people can be transformed if they can afford to pay a bribe to get a permanent
job (often in government service). The poor, like all of us, also like to invest in costly items that make
life more comfortable - better roofing, better furniture, a fan, a TV. One of these investment
opportunities - setting up a new business or expanding an existing one - has recently attracted a lot
of attention from the aid industry and from the new generation of banks that work with the poor. But
business investment is in fact just one of many needs and opportunities that require the poor to
become occasional ‘big spenders’.
3 Financial services for poor people
Defining financial services for poor people
In this essay we shall be concentrating on how the poor obtain the large lump sums they need from
time to time. We shall be reviewing the financial services - formal and informal - that have evolved
to serve this need. These are the services that are most urgent and frequent for the vast majority of
poor people, for the reasons set out in the previous section. They are the ones discussed in this essay.
Of course, there are other services that poor people use that are ‘financial’ in the wider sense, such
as those that ease the transmission or conversion of currency. Examples are sending money home
ix
from town or abroad. Apart from this brief mention, these services (important though they are to
many poor people) are not dealt with in this essay.
So, to return to our main question: how are the poor to get hold of the usefully large lump sums they
so often need? They might be lucky and have cash gifted to them, or be in some other way the
beneficiary of charity - but this can hardly be relied on. It is not a sustainable way of getting access
to large sums.
There are only three common ways of raising large sums:
The first is by selling assets they already hold (or expect to hold).
The second is by taking a loan by mortgaging (or ‘pawning’) those assets.
The third is by finding a way of turning their many small savings into large lump sums.
Let us review them.
Stocks and flows
The first method listed above - the sale of assets - is usually a straightforward matter that doesn't

ordinarily require any ‘financial services’. However, poor people sometimes sell, in advance, assets
that they don't hold now but expect to hold in the future. The most common example is the
advance sale of crops. These ‘advances’ are a form of financing, since the buyer provides, in effect,
a loan secured against the yet-to-be harvested crop. The advance may be spent on financing the
farming costs required to provide that crop. But they may just as likely be used on any of the other
needs and opportunities we reviewed in the previous section, or simply on surviving until harvest time.
The second method - mortgage and pawn - enables poor people to convert assets into cash and
back again. It is the chance (not always realised) to regain the asset that distinguishes this second
method from the first. As in the straightforward sale of assets, such services require the user to have a
stock of wealth in the form of an asset of some sort. They allow the user to exploit their ownership of
this stock of wealth by transforming it temporarily into cash. The most common examples are the
pawn shop in town and mortgaging land in the countryside.
These first two methods require that the users have assets, and poor people, almost by definition,
have very few assets. This fact severely limits the effectiveness of these two methods. It makes them
neither reliable nor sustainable. Only the third method is free of this limitation.
The third method enables poor people to convert their small savings into lump sums. This requires
the users to have a flow of savings, however small or irregular. It allows them to exploit their capacity
to make savings by offering a variety of mechanisms by which these savings can be transformed
into lump sums. The main mechanisms are
savings deposit, which allow a lump sum to be enjoyed in future in exchange for a series of
savings made now
loans, which allow a lump sum to be enjoyed now in exchange for a series of savings to be
made in the future (in the form of repayment instalments), and
insurance, which allows a lump sum to be enjoyed at some unspecified future time in
exchange for a series of savings made both now and in the future
Basic personal financial intermediation
The set of mechanisms associated with this third method needs a name that is less clumsy than
‘services which enable poor people to convert their small savings into usefully large lump sums’. I
suggest the term ‘basic personal financial intermediation’. I admit this is still a mouthful, but it does
describe the process at work here.

The process is one of ‘financial intermediation’ in the sense that a regular banker would recognise
1
,
because many small savings are ‘intermediated’ (‘carried across’) into lump sums. But the process is
‘personal’ because we are talking about how one poor person can turn her savings into a lump sum
for her own use (whereas bankers normally talk about intermediating the savings of many into loans
for a few - who may be entirely different people). Finally I call the process ‘basic’ because it is a
basic requirement of everyday life for most poor people.

1
The Economist defines a financial intermediary as ‘any individual or institution that mediates between savers
(that is sources of funds) and borrowers (that is users of funds).’ Pocket Finance, Economist Books, London,
1994, page 94.
x
Summary: financial services for poor people
Financial services for poor people are there to help them get hold of usefully large sums of cash when they need
the cash or have an opportunity to invest it
Assets (stocks) can be sold to raise cash, but this method is limited by the fact that the poor hold few assets.
Mortgaging or pawning assets – exchanging them temporarily for cash – is an important financial service for the
poor, but once again it is limited by the poors’ lack of assets.
The only reliable and sustainable way of raising lump sums of cash is to find a way of building them from your
capacity to save small amounts from time to time. I have called this method ‘basic personal financial
intermediation’.
Basic personal financial intermediation, or finding a way to convert a flow of savings into a lump sum, may
involve:
• a savings service that allows you to accumulate savings first and take the resulting lump sum later
• a loan service that allows you take the lump sum first as an advance against future savings
• an insurance service that allows you to take a lump sum at the time it is needed in exchange for a continuous
stream of savings
• or some combination of all three

Illustrations
Of all the propositions that I have put forward so far, the ones that people usually find most strange
are:
the idea that most poor people want to save, can save, and do save
the idea that loans are often nothing more than one way of turning savings into lump sums
The remainder of this first chapter is devoted to a small number of examples of ‘basic personal
financial intermediation’ that will, I hope, make these ideas feel less odd. Each example is a real
one that I have personally investigated by observing and talking to the people involved. Each
example, except the last, is typical of phenomena that are widespread among the poor all over the
developing world - though of course the detail will vary from place to place.
4 Deposit collectors
The need to find a safe place to keep savings is so strong that some poor people
willingly pay others to take their savings out of their hands and store them
We begin in India, in the slums of the south-eastern town of Vijayawada. There I found Jyothi doing
her rounds. Jyothi is a middle-aged semi-educated woman who makes her living as a peripatetic
(wandering) deposit collector. Her clients are slum dwellers, mostly women. Jyothi has, over the
years, built a good reputation as a safe pair of hands which can be trusted to take care of the
savings of her clients.
This is how she works. She gives each of her clients a simple card, divided into 220 boxes (eleven
rows and twenty columns), as shown here. Clients commit themselves to saving a certain amount
per box, in a certain period. For example, one
client may agree to save five rupees, at the rate
of one box a day. This means that at the end of
220 days (since there are 220 boxes) she will have
deposited 220 times 5 rupees, or 1,100 rupees
(that’s about $25 US). Having made this
agreement, it is now Jyothi’s duty to visit this client
each day to collect the five rupees. In the card
reproduced here the client has got as far as
saving 47 times, for a total of 235 rupees to date.

When the contract is fulfilled - when the client has
saved 5 rupees 220 times (which may actually
take more or less than 220 days, because slum
dwelling women are human beings and not slot
JYOTHI SAVINGSJYOTHI SAVINGS
5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5
5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5
5 5 5 5 5 5 5
xi
machines), the client takes her savings back. However, she doesn’t get it all back, since Jyothi needs
to be paid for the service she provides. These ‘fees’ vary, but in Jyothi’s case it is 20 out of the 220
boxes - or 100 rupees out of the 1,100 rupees saved up by the client in our example.
We can calculate Jyothi’s fee as a percentage of the cash she handles, in which case her fee, at
100 in 1,100 rupees, can be said to be 9%. Or we can look at it in another way and work out the
interest that her savers are earning on their savings. Obviously, since they get back less than they put
in, they are earning a negative interest rate, but what is that rate? Well, in our example the client has
saved 1,100 rupees over 220 days. That means that on average over the 220 day period she had
half that amount, or 550 rupees, on deposit with Jyothi. On that 550 rupees she has paid 100
rupees, or 18%. That was over a 220-day period, but interest percentages are best calculated at an
annual rate, so that it is easy to compare one rate with another. 18% over 220 days is the same as
30% over 365 days. So the ‘APR’ - the annual percentage rate - is about 30%. In other words, the
client is ‘earning’ interest at minus 30% APR.
Why should savers be prepared to accept a negative interest rate on savings? We can give two sorts
of answers, which complement each other. One sort of answer comes from economists. They would
say ‘these rates are so abnormal that there is obviously an imperfect market here’. They mean that
the demand for savings services is not being freely matched by the supply of savings services. That is
exactly correct in the Vijayawada slums. Apart from people like Jyothi, slum dwellers have very few
other places to put their savings. Banks are too remote, physically and socially, and don't like to
accept tiny deposits like 5 rupees a day. It is extremely hard to save at home, as we noted in the first
section. Competitors for Jyothi are few, perhaps because it takes a long time and a special sort of

person to build up the reputation for safety that Jyothi has.
The second sort of answer comes from the users of this system, and sheds light on the nature of their
‘demand’ for savings services. The first client I talked to was doing it to be able to buy school fees
and clothing for her two school-age children. She knew she had to have about 800 rupees in early
July, or she would miss out on getting her children into school. Her husband, a day labourer, could
not be relied on to come up with so much money at one time, and in any case he felt that looking
after the children’s education was her duty, not his. She knew she’d not be able to save so large an
amount at home - with so many other more immediate demands on the scarce cash she wouldn't
be able to maintain the discipline. I asked her if she understood that she was paying 30% a year for
the privilege of saving with Jyothi. She did, and still thought it a bargain. Without Jyothi, she wouldn't
be able to school the children. Other users told similar stories, and slum dwellers in a neighbouring
slum where there is no Jyothi at work actually envied Jyothi’s clients.
How does our concept of ‘basic personal financial intermediation’ help us to understand what is
going on here? Jyothi’s clients commit themselves to a series of equal and (more or less) regular but
tiny savings which Jyothi’s holds for them until they are transformed (intermediated) into a usefully
large lump sum (large enough to pay the school costs, for example). We can represent this
diagrammatically:
We can then improve our
diagram. We can label
the axes - time is on the
horizontal axis, and the
amount (or value) of
money is shown on the
vertical axis. We can show
Jyothi’s fee with a different
shading, as a deduction
from the lump sum. We
can show the savings and
the fees below the
horizontal axis, as

negative values, since the
client pays these in to
Jyothi, and the lump sum
above the axis, as a
a long series of small savings… and a usefully large lump
CHART ONE:
BASIC PERSONAL FINANCIAL INTERMEDIATION
xii
positive value, since the client gets this out of the system. Our diagram will now look like this, and a
similar convention will be used in all subsequent diagrams:
Note one more thing
about services like Jyothi’s:
that clients often start a
second cycle as soon as
the first cycle is finished.
That is why I have drawn a
broken vertical line after
the pay-out, to indicate
the end of one cycle and
the beginning of the next.
Some of you will be
thinking ‘this is an
extraordinarily elaborate
way of describing
something as simple as
saving up money and
then withdrawing it’. I can
only ask you to remain patient, because I hope that getting the diagram right now will help us
explain, later, other less simple phenomena in the world of financial services for poor people.
valuevalue

CHART TWO:
JYOTHI SAVINGS
value
time >
pay-ins
pay-out
next cycle starts >
xiii
Summing up Jyothi savings:
The market for savings deposit in slums is ‘imperfect’ (demand is not matched by supply).
Slum dwellers want to turn their savings into lump sums for many different needs and opportunities.
Unable to save at home, and unable to go to remote unfriendly banks, they trust their savings to unlicensed
informal peripatetic collectors.
When they find one that they can trust, time and time again, they are willing to pay a high price (as much as
30% a year) to have that collector take away their savings and store them safely until needed.
The service that these deposit collectors render represents the most simple version of ‘basic personal financial
intermediation’ for poor people.
We look next at one of Jyothi’s competitor - the urban moneylender.
5 The Urban Moneylender
In an environment where the demand for savings services far outstrips supply, it is
not surprising that many loans to poor people turn out to be just another way of
turning savings into lump sums.
There are many kinds of moneylender. Among them there is one kind that is common in many
urban slums of the sort where deposit collectors like Jyothi work. Indeed, I have taken my example
from Vijayawada again because I want to draw a comparison with Jyothi. I got to know this
moneylender’s clients in a slum not far from the one where Jyothi works.
His working method is simple. He gives loans to poor people without any security (or ‘collateral’), and
then takes back his money in regular instalments over the next few weeks or months. He charges for
this service by deducting a percentage (in his case 15%) of the value of the loan at the time he
disburses it. One of his clients reported the deal to me as follows.

‘I run a very small shop’ (it’s a small timber box on stilts on the sidewalk inside which he squats and
sells a few basic household goods) ‘and I need the moneylender to help me maintain my stock of
goods. I borrow 1,000 rupees from him from which he deducts 150 as interest. He then visits me
weekly and I repay the 1,000 rupees over ten weeks, at 100 rupees a week. As soon as I have paid
him off he normally lets me have another loan.’
This client - Ramalu - showed me the scruffy bit of card which the moneylender had given him and
on which his weekly
repayments are recorded.
It was quite like the cards
Jyothi hands out. There are
many other similarities
between Jyothi and the
moneylender. We can see
that if we redraw our
diagram to show this
moneylender’s system:
The main difference - the
fact that the pay-out
comes first, as a loan, is
immediately apparent. But
let us look at the
similarities. In each case
the client is using the
service to swap a series of small regular pay-ins (or savings) for a usefully big pay-out. In other words,
these are both forms of basic personal financial intermediation. With the urban moneylender, the
pay-out comes first, and can be understood as an advance against future savings. Indeed, very
many loans to poor people are actually advances against savings, as we shall see.
Another similarity is that clients often proceed straight into a second cycle - and then a third and so
on. When you have done several cycles it can hardly make much difference whether in the first
cycle the loan or the savings came first - you may not even be able to remember. You have got into

a rhythm. Every day (or week or month or whatever) you make a small pay-in and every now and
CHART THREE:
URBAN MONEYLENDER
advance (loan)
pay-ins (repayments)
fee (interest charged)
next cycle starts >
xiv
then (every 220 days or every ten weeks or whatever) you get a usefully big pay-out. Just what you
need if you’re poor, as we saw earlier in the chapter. This is the essence of basic personal financial
intermediation.
As in all cases of basic personal financial intermediation, the size of the pay-out is directly linked to
the size of the pay-ins. In the case of Jyothi, the client makes the decision, by choosing the size of
the pay-in. In the case of the moneylender, the moneylender makes the decision, by choosing the
size of the loan (or at least its maximum size). To do this, he has to judge the client’s capacity to
save, and in this he is often helped by a history of previous similar deals with the same client or with
people in similar situations.
This brings us to another important difference, the difference in price of the two services. The
moneylender is more expensive. Calculating his rates in the same way we calculated Jyothi’s, we
can see that the moneylender charges 15% of the cash he handles (as opposed to Jyothi’s 9%),
and charges an APR of around 180%
2
(as opposed to Jyothi’s 30%).
The client pays the moneylender more but of course the client gets more for his (or her) money. For
one thing, the moneylender accepts the risk of the client’s being unable or unwilling to make the
pay-ins, a risk which Jyoti doesn't face (indeed, her clients have to accept the risk that she’ll run off
with their money). Secondly, the moneylender puts up the initial finance for the first cycle, whereas
Jyothi needs no capital to run her business. Thirdly, the moneylender has to use his judgement about
the size of each contract, while Jyothi can happily leave that to her clients. For all these reasons
clients pay the moneylender more than Jyothi for an essentially similar basic personal financial

intermediation service. We can now see why the women in the slum next to Jyothi’s envied Jyothi’s
clients their access to a safe and relatively cheap way to build a lump sum from their savings.
6 The Merry-go-round
But both sets of clients - Jyothi’s and the moneylender’s - could run the same sort of service for
themselves, for free. To see how, we need to look at ROtating Savings and Credit Associations, or
ROSCAs. Since there are many kinds of ROSCA we’ll look at a very simple one in this chapter, the
‘merry-go-round’ as it is practised in the slums of Nairobi, Kenya.
Mary, a woman whose ROSCA I studied there, is, like Ramalu, a very small vendor. She sells
vegetables from a shelf set in the window of her hut. She is a member of a merry-go-round that has
fifteen members, including herself. This is what they do.
Every day, day-in day-out, each of them saves 100 shillings. So each day a total of 1,500 shillings
(about $40) is saved. Each day one of the fifteen women takes the full 1,500 shillings. After each of
the fifteen women has taken the ‘prize’ in turn - which takes fifteen days of course - the cycle starts
again. Mary was ‘serial number 7’ in the cycle. So seven days after the start of the first cycle, and
then every fifteen days, she gets 1,500 shillings in return for putting in 100 shillings each and every
day. Mary told me she had been in this merry-go-round with the same fellow-members for two and
half years.
Here is Mary’s merry-go-round pictured in our usual diagram, where the ‘basic personal financial
intermediation’ function and its relationship to Jyothi and the moneylender is, I hope, clear.
The ‘do-it-yourself’ nature of
this device gives it its
particular advantage over
the other examples. There
are no fees or interest
payments. You get back
exactly what you put in. Of
course, there are other,
non-monetary costs. Mary
and her friends have to
organise it and maintain

trust and agree among

2
For ten weeks the client had an average loan in his pocket of 425 rupees. On that he paid 150 rupees interest,
or about 35%. That was for ten weeks. At a yearly rate (52 weeks) we divide 35% by 10 and multiply by 52 to
arrive at 180%.
CHART FOUR:
MERRY-GO-ROUND
pay-out
pay-ins
next cycle
begins
xv
themselves the number of members and the size and frequency of the pay-ins, tasks that are not
needed if you use a commercial provider like Jyothi or the moneylender.
Mary takes her merry-go-round very seriously. The total value of the stock of her ‘shop’ is only a little
over 1,500 shillings. So when she has to dig into her working capital to pay for extra costs for her two
children (Mary has no husband) she can do so safe in the knowledge that provided she is faithful to
her merry-go-round she’ll get a 1,500 shilling lump sum within the next fortnight, and can then re-
capitalise her shop. She once tried joining an ‘NGO’ that offered a bigger loan, but she found that its
repayment schedule was too long to suit her needs, so she left. Instead, she joined another, longer-
period ROSCA which she uses to build up her savings over a longer term, for use in schooling her
boys.
Many ROSCA members in Nairobi join more than one ROSCA. This helps them get round a
disadvantage of ROSCAs - an inflexibility in which everyone has to save the same amount in the
same period, whereas individual households may have actual needs that vary in quantity and date.
7 Rabeya’s ‘Fund’
Is it possible to devise a type of ‘basic personal financial intermediation’ device that includes most of
the advantages and eliminates most of the disadvantages of deposit collectors, moneylenders and
ROSCAs? The last two examples in this chapter show two attempts to do exactly that.

We start with what the slum-dwellers of Dhaka, in Bangladesh, call a ‘Fund’. This is a type of savings
club that can be found all over the world. In many places, including Dhaka, it is the main alternative
to the ROSCA among user-owned devices for basic personal financial intermediation. It differs from
a ROSCA in that the savings that its members deposit accumulate in a ‘fund’ from which members
may borrow – but only if they wish to. Here is our diagram for a Fund:
The diagram is beginning to get more complicated, and this reflects one of the disadvantages of
Funds - they require more deliberate and careful management to make them work well. This is how
they work, based on what happened to a Fund that I tracked for some months in Dhaka in 1996:
In that Fund there were 23 members, and all of them had committed themselves to save on a
weekly basis for one year, after which the Fund was to close. Each member chose how much to
save, but it was always some multiple of 10 taka
3
. In practice, some saved 10, some 20 and a few
were saving 50 taka a week (shown below the line in the diagram as ‘weekly pay-ins’). As they came
in, these savings were stored with Rabeya, the Fund’s chairperson, a housewife who had run many

3
There are about 50 Bangladeshi taka to one US dollar
CHART FIVE:
RABEYA'S FUND
optional
advance
pay-ins
extra pay-ins to repay optional
advance
fee (interest paid on optional
advance)
profit
(interest
earned)

pay-out
next
cycle
may
begin
xvi
Funds in her neighbourhood, where she was well known. She kept a simple set of accounts in a
school exercise book.
As soon as this cash on hand became big enough, members with a need or an opportunity for a
lump sum were allowed to borrow from it (shown above the line in the diagram as ‘optional
advance’). The terms of these loans were straightforward – the borrower had to pay interest of 5% a
month, and had to repay the loan before the end of the year (repayments and interest on loans are
shown below the line in the diagram). Decisions on who took a loan and how much they took were,
according to the Fund’s rules, made by the members collectively: in practice the Chairperson had
by far the biggest say. She strove to make sure that everyone who wanted a loan could get one,
and that no member borrowed an amount that was beyond what the she estimated they could
repay in the time allowed.
At the end of the year, the total fund, including the interest earned on the loans made from it, was
put on the table and shared by the members in proportion to the savings they had made. It worked
out that for each 10 taka saved per week, members got back 580 taka (shown as ‘pay-out and
profit’ in the diagram). Thus a member saving 10 taka a week saved 520 in the year (52 weeks) and
got a ‘profit’ of 60 taka. Members got this profit on their savings irrespective of whether they took a
loan.
How did this Fund perform in comparison with the other devices? Compared to the moneylender the
Fund’s advances are - at 60% APR
4
, instead of 180% - a much cheaper way of borrowing lump
sums. As a way of saving up, the Fund is not only cheaper than saving with Jyothi, it returns a good
profit. You earn 11.5% over and above what you put in (instead of losing 9% as with Jyothi)
5

. This is
an APR of plus 23% (instead of minus 30% with Jyothi)
6
. As with Jyothi you can choose how much to
save each week. But you don't get a daily visit, and you can’t choose your own start date, since that
is a decision that has to be made collectively. Best of all, the Fund offers you two ways of swapping
small pay-ins for lump sums, instead of one. You save up and withdraw, but if you wish you can take
an advance as well. This double opportunity makes the Fund more flexible than the ROSCA - at the
cost of more paperwork and management.
This added management burden makes Funds less transparent and so more vulnerable to fraud
than ROSCAs. Some conditions help to minimise this risk. For example, where Funds are very
common the ratios they use tend to converge – so that almost every Fund in the area charges the
same rate for loans and guarantees the same minimum return on savings. This ‘institutionalisation’ of
Funds makes it easier for poor illiterate people to know exactly what they’re getting themselves in to.
Not all Funds are time-bound in the way that Rabeya’s was. Some go on for an indefinite period. But
being time-bound is a very healthy feature that good Funds share with ROSCAs. During a ROSCA or
at the end of a time-bound Fund either you get your money back or you don't. If their ROSCAs or
Funds don't produce the goods, the members walk away and the device dies. As a result, poor
managers are soon out of a job, and members flock to others with a sound record. This makes sure
that the vast majority of such savings clubs are well run. I call this an ‘action audit’ and it substitutes
very well for the sort of formal but less easily understandable audit that professional savings banks get
accountants to do.
Funds can be wholly user-owned (run by the people that use them, as in Rabeya’s case), or run by
club officers on behalf of users, as when a church or social club runs them. They can also be run
professionally, and some bigger church and trade-association Funds are more or less ‘commercial’

4
5% a month is equivalent to 60% a year. The formal equation for calculating APR is - professionals will note -
different from the simplified (but useful) calculation I use. The law (in the UK) requires the use of the formula (1
plus the interest rate for the period quoted) to the power of the number of such periods in a year, minus 1. Under

this formula 5% a month is an APR of 79.5%. (1 + 0.04)
12
-1), not 60%. This allows for the fact that if you pay the
interest each month instead of in one lump at the end then you are out of pocket and the loan has effectively cost
you more. This extra cost can be significant in loans on which interest is paid at short intervals, as in home
mortgages in the rich world. My calculation ignores this sophistication, though where interest is paid at the end -
as in the calculation in the previous footnote - there is no difference.
5
If you are saving 10 taka a week you put in 520 taka in the year and earn and extra 60 rupees. 60 is 11.5% of
520.
6
If you are saving 10 taka a week then over the year you have an average of 260 taka on deposit. On this you
earn 60 rupees. 60 is 23% of 260.
xvii
in that what they charge for the service generates a surplus, ensuring their continuity. We shall look at
some examples in the third chapter. Meanwhile, our last example in this chapter is also run
commercially, and is a deliberate attempt to sum up many of the lessons of basic personal financial
intermediation in one device.
8 SafeSave
If we return to the early sections of this chapter and review, in their light, the examples shown so far
(Jyothi, the moneylender, the ROSCA and the Fund), we shall find two respects in which the
circumstances and needs of the poor are still not being met.
First, we noted that poor people need to store savings for the long run, for widowhood or old age or
for their heirs. None of the examples shown so far helps them to do this (or at least not directly).
Second, we noted that poor people’s ability to save fluctuates with time, so that they may be able to
save a lot in one week and very little in another. But in all our examples so far there is the requirement
for a (more or less) fixed saving at a fixed interval (the same sum each day for each box on Jyothi’s
card or for Mary’s ROSCA, or for each week for the moneylender or for Rabeya’s Fund).
Both of these shortcomings are particularly difficult for the very poor. It is the very poor who suffer
most hardship in old age and most need financial protection for the end of their lives. And many

poor people get excluded from these devices - and often indeed exclude themselves - out of
anxiety that they won’t be able to save the same amount every day (or week, or month) for a whole
year (or other period).
The SafeSave diagram shows how SafeSave tries to get round these shortcomings.
SafeSave has Collectors (field staff) who visit each client each day at their home or workplace. They
provide the same opportunity to save (or repay) that Jyothi and daily ROSCAs do. On each
occasion, clients may save, but in any amount they like, including zero. The ‘pay-ins’ in the diagram
show this - they vary over time. From this accumulation of savings clients may withdraw a lump sum
at any time they like - this is shown by the solid black amounts. Then, as in a Fund, they can take
optional advances, but - better than an Fund - clients repay when they like and can take as long as
they like as long as they pay the interest (shown as fees) each month. Finally, as in a Fund, they get a
pay-out of their accumulated savings plus profits. But unlike in a time-bound Fund like Rabeya’s, they
can leave these savings on deposit for as long as they like and earn even more profit the longer they
leave them in. The only respect in which this flexibility is compromised is that they cannot withdraw
CHART SIX:
SAFESAVE
pay-ins
pay-out
profit
withdrawals of pay-ins
optional advance
repayments on advance
interest on advance
xviii
from savings while they are holding an advance (except to repay the advance) and for this reason
clients are allowed to hold more than one account.
The current version of SafeSave running in the slums of Dhaka, Bangladesh pays clients a little under
10% a year on savings (much less than Rabeya’s Fund but more than formal banks) and charges an
APR of about 28% on advances (much less than Rabeya’s Fund but more than formal banks).
SafeSave raises many questions. One of them is whether SafeSave can be run profitably, generating

surpluses that guarantee its sustainable and fuel its expansion. Unless that can be shown to be the
case, SafeSave will not contribute much to banking with the poor. Though early signs are
encouraging, SafeSave is still young - it began only in 1996 - and more time is needed to see if it will
pass this crucial test. No more will be said on this issue here, since this chapter is concerned with
‘basic personal financial intermediation’ and focuses on the user’s perspective.
The main question raised by SafeSave in that context is discipline.
We have already seen that without discipline it is hard to save. This is true whether those savings are
made following an advance against savings (as with moneylenders), or whether they precede a
withdrawal or advance (as with a deposit collector like Jyothi) or are made both before and after a
withdrawal/advance (as in a ROSCA). Moneylenders enforce discipline by their regular weekly visit,
and Jyothi does it by daily appearances on the doorsteps of her clients. ROSCAs fail if their self-
imposed discipline falters. SafeSave is no different, except that it has given up some things that
undeniably promote discipline very strongly - uniformity of deposit size, and regularity of deposit. In
all the other examples shown so far the user pays a set amount at a set interval. In SafeSave the user
may pay at any interval and in any amount - including zero.
The risk is, therefore, that without any compulsion to pay a set sum at a set interval, SafeSave’s clients
will simply fail to save. SafeSave’s experimental aspect is precisely that it is testing the extent to which
a frequent and reliable opportunity to save is a way of maintaining savings discipline. So far, the
indications are good. It looks as if the frequent opportunity to save - having someone knock on your
door each day - is as good, or even better, as a way of maximising savings, as the obligation to pay
a set sum at a set interval.
Summing up SafeSave:
SafeSave is a deliberate attempt to set up a financial service scheme for the poor which meets their
circumstances and needs as understood by this author over twenty years of research and practice.
It allows for the fact that the poor can save and want to save - but can save only in small (but variable sized)
amounts and can’t save each and every day.
It allows for the fact that the poor need to turn those savings into usefully large lump sums at both short and
long-term notice, and sometimes without notice. It recognises that to help them do this it must allow them - on a
daily basis - all three of the ‘basic personal financial intermediation’ functions:
• the chance to save and withdraw

• the chance to take an advance against future savings
• the opportunity to store up savings for long-term needs
SafeSave recognises that no-one can save without discipline, and offers a daily opportunity to save to all its
clients as a way of developing and maintaining that discipline.
SafeSave is thus the most flexible of all the examples dealt with in this chapter, and because of this is the most
attractive to the very poor who can be frightened off by the need to pay set sums at set intervals.
It may occur to you that - with the exception of the doorstep service offered by its Collectors - the financial
services SafeSave offers are rather like what is available over the counter to ordinary customers of banks in the
rich world. It is a combination of current account, savings account, long-term deposit, and loans.
Should that surprise you?
Conclusion
This chapter has introduced ‘basic personal financial intermediation’ - the process through which
people turn their savings into usefully large lump sums of money. Poor people need this process as
much as anyone else, because poor people can save and poor people have frequent need,
throughout their lives, of ‘usefully large lump sums of money’. Other ways of getting hold of large
sums of cash, such as being the beneficiary of charity, or selling or pawning assets, are either
unreliable or unsustainable. The task of financial services for the poor, therefore, is to deliver them
mechanisms through which the swap from savings into lump sums can be made.
xix
As an introduction to the wide variety of such mechanisms, the chapter has described three informal
devices. Deposit collectors will accept people's savings and return a lump sum to them,
moneylenders will provide the lump sum up front and then collect savings in repayment, and
ROSCAs allow people get together to make savings from which each in turn takes their lump sum.
Elements from these three systems can be combined to provide a more flexible service, as we saw
in the example of Rabeya’s Fund.
These devices are all time-bound, but poor people's needs for basic personal financial
intermediation are never-ending, so many poor people engage in cycle after cycle with their
deposit collector, money lender, ROSCA, or Fund. SafeSave, the last example in the chapter,
illustrates one way of serving poor people's longer term needs for swaps, by allowing them to keep
money on deposit for the long term. SafeSave, unlike the other devices discussed so far, allows

savings deposits to be made as and when the saver has them to hand: the idea behind this flexibility
is that the very poor, who may feel unable to save set sums at set intervals, can also avail the
service.
All the main ideas of my essay have now been expressed. If you wish to read on, you will find, in
Chapters Two and Three, much more detailed descriptions of the kinds of devices that you can
expect to find in slums and village in the developing world. After that, in Chapter Four, I describe a
little of the recent work that has been done by outsiders to bring more and better financial services
to the poor.
xx
Notes
Town and country. We noted that people can save money when it is ‘on the way out’ (during
expenditure) as well as when it is on the way in (at the time income is received). This helps to correct
a common misapprehension about the differences between town and country. I sometimes hear it
said that ‘in the urban slums people can save because they have a variety of sources continually
producing income - but rural farmers may only get income at the end of each growing season, and
that is the only time they can save’. This ignores the fact that in many countries the rural poor are
often not farmers, having lost their land. They are day labourers and may earn on a daily or weekly
basis. But even those poor who are farmers go to market frequently - once or twice a month, or
even weekly - to buy perishable or expendable items like salt, fresh food, kerosene oil, matches, and
so on. The money they use for this can come from several sources, including the sale of short-term
farm produce like eggs, chickens, or fruit, or from income from supplementary work like cutting
firewood, or from selling bigger items in which they have stored (or saved) value, such as stocks of
grain, pigs or goats. Each such market visit presents an opportunity to save some money, even if this
saving simply converts a non-money form of saving (the piglet) into cash savings.
Saving in kind. Mentioning piglets reminds us that poor people often save in non-money ways. These
non-monetary savings may be very important to their owners, but they are not the subject of this
essay, except in the following sense. Poor people sometimes store their savings in livestock or other
non-money ways simply because they haven’t got access to a safe, rewarding, inflation-proof place
to save money. Once they are given the opportunity, poor people often choose to convert some of
their non-money savings into cash savings. This is because cash savings can be more useful, and

less risky, than non-money savings. The piglet may get sick and die or be stolen, and if you all need
is two dollars to buy medicine for a sick child, it is rather troublesome to have to sell a piglet worth
thirty dollars
7
. Also, non-money savings are themselves easier to manage if you have access to a
cash-savings service. After all, when you’ve sold the thirty-dollar piglet you need somewhere to put
the twenty-eight dollars left over after buying the medicine. And if you save in the form of gold
ornaments, as some poor people do, how did you save up the cash to buy the ornament in the first
place? SafeSave customers often use SafeSave to save up enough to buy an ear-ring. The
inescapable conclusion is this - that a cash-savings service is useful even to people who prefer to
store most of their savings in non-cash forms. As the world becomes ever more monetized many
poor people are coming to see that for themselves, and the demand for financial services grows.
Pawning. In some countries pawnshops have been outlawed, sometimes so successfully that some
readers from those countries require an explanation of pawning (after which they normally recognise
the phenomenon which tends to exist in their ‘grey’ economies under some local name). A pawn is
a movable asset (most commonly a precious metal, above all gold) that is taken as security for a
loan by a lender - the ‘pawnbroker’
8
. You take your gold ring along to him and he weighs it and gives
you, if you’re lucky, about 60% of its market value. When you pay him back (with interest) you get the
ring back. If you never pay him back he keeps the ring and in the end sells it. Pawning is to the town
what mortgaging land is to the countryside - an example of a class of financial services for the poor
by which assets can be turned into cash and back again.
Other ways to get hold of usefully large sums of money. We noted that you can sometimes sell
assets that you expect to hold in the future - selling your chickens before they’ve hatched, as it were.
As well as selling assets like crops in advance, you can also sell your labour (or that of your children or
spouse) in advance. This is common in rural Bangladesh and in several other countries. We could list
other examples of ingenious ways to get hold of money, but this essay sticks to those that are

7

Money is ‘fungible’ - it can be quickly converted into services or goods (including medicine and piglets). It is the
point of money to be fungible. That’s why we invented it.
8
This definition of pawning is similar to the conventional one for mortgaging. Another way to look at pawning is to
say that it is not part of a loan contract, but is the sale of goods linked to a promise to buy it back again, under
which part of its value is forfeited if the re-purchase fails to take place.
xxi
common everywhere, and which involve mainly financial transactions, rather than sales of goods or
labour.
Chapter Two:
Doing it yourself: ROSCAs and ASCAs
Savings clubs
Savings clubs are groups of people who come together to set up and run their own
basic personal financial intermediation services. There are two kinds of clubs - the
ROSCA kind (where everyone puts in and takes out the same amount) and the
‘accumulating’ kind (where they don't).
The world of money management for the poor is rich and complex. Schemes and services have
long histories, and countless variations have evolved. Geographic areas have come up with
solutions tailored to their particular social and economic conditions. As a result, it’s not easy to
categorise financial services for the poor. Nevertheless, this chapter and the next divide the services
into three classes - savings clubs, managers and informal providers - a classification based on who
owns and manages the services. The categories are robust enough to be useful, if not water-tight.
Savings clubs are composed of people who come together to pool their savings in various ways.
These clubs are owned and managed by their members, and it is this characteristic that
distinguishes savings clubs from the other two classes. There are, however, two main kinds of user-
owned savings club. There is the ROSCA kind where the cash rotates evenly between all the group
members (as in Mary’s merry-go-round), and an ‘accumulating’ type where some members borrow
and others don’t (as in Rabeya’s Fund). For the accumulating (Fund) type I am going to use the
name that Fritz Bouman gave them – the ASCA, for Accumulating Savings and Credit Association
9

.
Managers are those that run savings clubs for other people. Religious and welfare organisations
often do this on a voluntary or non-profit basis, but there are also commercial managers, such as
those who earn a fee for managing ROSCAs for people – I call them ‘chit managers’, after the Indian
name for the ROSCA.
Informal providers are a mixed bunch who have in common the fact that they provide basic
personal financial intermediation services to others. Deposit collectors (such as Jyothi), moneylenders
(like Ramalu’s) and pawnbrokers are examples. Usually they deal with users of their services on an
individual basis, and most charge for their services.
My classification system therefore looks like this:
Three classes of basic personal financial intermediation services for the poor:Three classes of basic personal financial intermediation services for the poor:
1. Savings Clubs (owner-managed) 2. Managers 3. Providers
(described in this chapter) (described in the next chapter)
ROSCAs (where the
cash rotates evenly
between members)
ASCAs (where it
doesn’t)
Including religious and
welfare organisations,
and ‘chit’ managers
Including deposit
collectors, moneylenders,
and pawnbrokers
and this gives us a structure for this and the following chapter. This chapter deals with ROSCAs and
ASCAs, while the next describes Managers and Providers. A section on the ingenious ‘ubbu-
tungnguls’ of northern Philippines, is included at the end of this chapter as a demonstration of the
inventiveness of poor people when it comes to managing money – and as a reminder of how hard
it can be to categorise their inventions.

9
I used ‘Funds’ to describe Rabeya’s savings club in Chapter One because that’s what their users call them.
xxii
1 The ROSCA
The ROSCA is the world’s most efficient and cheapest financial intermediary
device. The best form of ROSCA – the auction ROSCA – matches savers perfectly
with borrowers, and rewards both of them.
With its description of Mary’s savings club or ‘merry-go-round’ Chapter One provided an example of
how the poor can and do get together to manage their own basic personal financial inter-
mediation. The merry-go-round is just one of many variations of the ROSCA, or rotating savings and
credit association. ROSCAs are found in their tens of thousands on every continent, and have been
for many years. There are references to ROSCAs in Japan dating back six hundred years
10
. This essay
is not going to tell the history of the ROSCA, nor will it offer evidence about the huge numbers of
ROSCAs found round the world, since there is documentation already available, as the bibliography
shows. Rather, noting that the ROSCA is indeed an extremely popular intermediation device, this
essay will try to honour it by describing as simply as possible its major variants and explaining their
differences.
Definitions
It was the anthropologist Shirley Ardener who devised what has become the standard definition of a
ROSCA:
an association formed upon a core of participants who make regular contributions to a fund
which is given, in whole or in part, to each contributor in rotation
Thus, in Mary’s merry-go-round, there are fifteen members (the ‘core of participants’) each of
whom makes a daily contribution of 100 shillings. That daily total (1,500 shillings) is given in
whole to each contributor in turn. The process takes fifteen days.
In what follows I use the word ‘round’ to refer to each distribution of the lump sum (the
number of which will equal the number of members). The word ‘cycle’ is used for the
complete set of rounds, after which the ROSCA comes naturally to an end (though it may be

repeated, with or without variations in the number of members, or in the amount and
frequency of the contributions). In Mary’s case there is a round each day for a fifteen-day
cycle, and then they start another cycle.
The ROSCA’s advantages….
The very elegance and neatness of the ROSCA gives it great appeal, and like many others I’m drawn
to it partly for that reason. I joined a ROSCA in Mexico in 1974 and have been fascinated by them
ever since. So they get first place in this chapter.
The virtues of ROSCAs are apparent in Mary’s club, which neatly arranges the small daily savings of
fifteen people into a series of fifteen large lump sums, which each member in turn enjoys. The
ROSCA, which then ends (only to be reborn into another cycle), has cost no money to run and is
wonderfully transparent - without elaborate books its accounts are clear to each and every member,
even if they include the illiterate. No outsiders are involved, no one is beholden to anyone else, and
no one has profited from anyone else’s difficulties. Moreover, no money has had to be stored by the
managers of the ROSCA, because all cash passes from members to member directly. This has two
healthy results. Firstly, it greatly reduces the risk of misappropriation. Secondly, it makes ROSCAs
extremely efficient. Indeed, ROSCAs could reasonably claim to be the most efficient intermediation
device around, since at each round the savings of many are transformed instantaneously, with no
middlemen, into a lump sum for one person.
…and its perceived disadvantages
However, when people first hear about ROSCAs they often react by listing their disadvantages - as
they see them. Usually, their first objection is ‘what stops those who first get the lump sums from
running away?’. The next is ‘but the system is unfair - the ones who get the lump sum first have a
huge advantage. They get an interest-free loan at their fellow-members’ expense’.

10
That is, before modern banking evolved in southern Europe.
xxiii
We have already hinted at the answers to these two objections, in the first chapter. People like to
save regularly if they can, to build up lump sums, so even the ‘end-takers’ still benefit from a ROSCA
compared to paying a deposit-collector like Jyothi or a moneylender. And people tend not to run

away from services that they like. However, we shall be able to build even better answers to these
objections by looking at the ROSCA in more detail.
Four ways of running ROSCAs
We start by listing the four main ways in which ROSCA users decide the order in which the lump sum
is taken. They are:
1. by prior agreement
2. by agreement each round
3. by lottery
4. by bidding for the lump sum
Deciding the order of the draw by prior agreement
Mary’s merry-go-round falls into type 1. This type is particularly appropriate when the intention is to run
many cycles of the ROSCA one after the other, as in Mary’s case. After a few cycles, any ‘unfairness’
in the order has shrunk to insignificance, and every member’s situation is the same - she gets her
lump sum every fifteen days (for example). This pattern of prior-agreement multi-cycle ROSCAs is the
dominant form of ROSCA in Nairobi’s slums. It provides slum-dwellers with a secure way of saving
regularly and continuously. Its simplicity - no decisions about the order of disbursement need be
taken apart from the initial one, and no mechanism like lotteries or auctions are required - suits this
continuous, routine savings function especially well. It means that members don't have to get
together in a meeting each time the lump sum is taken, and many such ROSCAs run without
meetings, or hold meetings only at the close of each cycle (which may also be the start of the next).
Very convenient.
Deciding the order of the draw at each round
Where members are well acquainted with one another ROSCAs sometimes function as type 2, with a
fresh decision about who gets the lump sum made at the time of each round, usually on the basis
of who needs it the most. There are probably fewer of this kind of ROSCA than of any other kind,
perhaps because of the difficulties of assessing ‘need’ without recourse to the price mechanism (see
auction ROSCAs below), and the risk that the more articulate or the more cunning will manipulate
the process.
But there is a variant of this type that is quite common, in which the ROSCA is initiated by someone
who suddenly needs a lump sum and who gets friends to join in. Thus in the mountainous north of

The Philippines I have met rural schoolteachers who go for many months without running a ROSCA,
until one of them wants cash to furnish a new home and calls on her fellow teachers to start a
ROSCA (usually funded from monthly salaries). She takes the first lump sum, and accepts
responsibility for the management of subsequent rounds until the ROSCA finishes
11
. Some lottery and
auction ROSCAs (see below) are also started in this way by an individual with a pressing need.
Deciding the order of the draw by chance – lottery ROSCAs
‘Lottery’ ROSCAs are a huge and varied class of ROSCA found almost everywhere. In some countries
they dominate - Bangladesh is an example. The lottery avoids the problems of any perceived
‘unfairness’ in the order in which the lump sum is taken, or of comparing people’s needs, by leaving
that order to chance. Typically, names are drawn out of ‘hats’ (or the local equivalent). Every
member’s name goes into the ‘hat’ in the first round, but winners are excluded from the lotteries of
subsequent rounds. Obviously for the last round no lottery is needed, there being by then only one
remaining member who hasn’t yet received the lump sum.
The lottery itself also generates a certain amount of excitement, which brings a crowd of onlookers
which in turn helps to make the process public and fair - though this ‘festival air’ tends to die down
after a while. And of course members sometimes find ways round the arbitrariness of the lottery.

11
Such patterns of reciprocal obligation characterise many other cash exchanges that are not strictly speaking
‘clubs’. Details of arrangements such as the neota of northern India, in which families are duty-bound to
contribute cash for weddings among their neighbours, and then expect to receive the same help when they have
a wedding, are reported in Jodkha and are summarised in Rutherford [1996, 1]
xxiv
Friends may agree to ‘swap’ (or share) their luck where one has a more pressing need than another,
or one member may even ‘buy’ another member’s lucky draw.
Precisely because lottery ROSCAs don't involve a group of friends deciding which among them most
needs the cash, they can afford to have a more varied membership made up of people who don't
know each other very well, or who are complete strangers. In Bangladesh, a typical ROSCA in the

capital, Dhaka, is run by a small-time shopkeeper who arranges the regular lottery. Not everyone
comes to the meeting, and many members pay as and when they can, often between meetings,
sometimes in instalments. The shopkeeper keeps the records of who has paid, and chases up late-
payers. In the ‘moral economy’ of Dhaka it is not yet considered proper for such ‘managers’ to run
ROSCAs commercially, so he (or she) bashfully accepts ‘tips’ from members as a reward for this work.
Deciding the order of the draw by bidding – auction ROSCAs
By leaving the selection of ‘winners’ to chance, lottery ROSCAs are more flexible and less
troublesome, and can cater to a wider variety of people and of needs than where the winner is
decided round-by-round by group consensus. As we saw, however, members sometimes ‘buy’ a
lucky draw from another member. But there is an even more flexible way to cater fairly to a wide
range of people and their individual needs, and that is by setting up a market to decide who should
take the lump sum at each round. This allocates cash to the member who most values it at the time,
while compensating others by rewarding them richly for their patience. It thus benefits both
‘borrowers’ and ‘savers’, and elegantly arranges them in serial order with those who most need to
borrow taking the lump sum at the beginning and those most content to save taking it at the end.
This is how they work. Imagine a twelve-person ROSCA that meets monthly with each member
contributing $10 (that is twelve ‘rounds’ for a twelve month ‘cycle’). At each round $120 is available
as the lump sum. At the first round those members in immediate need of cash choose to bid for the
lump sum. Let us say that five members want the money, but the one who most wants it is willing to
bid $24, and wins. She then takes $96 of the lump sum ($120 minus 24), while her bid of $24 is given
back, in equal shares, to each of the twelve members
12
, who walk off with $2 each (thus making a
net contribution of only $8 that month).
As the rounds proceed, the size of the winning bid tends to diminish, since there are fewer and fewer
people in the auction. This is so because, as in other ROSCAs, each member takes the lump sum -
or a part of it - once only. At the last round there is no need of an auction, because there is only one
member left in. He gets the full $120.
Doing the sums for auction ROSCAs
Calculating how each member fares in such a ROSCA has caused arithmetic mayhem among the

experts, so let’s make some simplifying assumptions. Let’s assume that the members who won the
first four rounds all bid $24, members taking rounds five through eight each bid $12, and in the last
four rounds there were no bids at all, so those members got the full $120. The bids total $144 (four
times $24 plus four times $12). These bids were redistributed equally among the members, so each
member got $12 back ($144 divided by 12). The total amount contributed by each member must
equal 12 rounds of $10 each, which comes to $120, less the $12 from their share of the bids, for a
total of $108. Contributions are thus the same for each and every member. But the total amount
taken out by each member varies. For example, the first member took out $96 on the first round,
while the last member took out the full $120 but had to wait until the last round.
Let’s look now at that last member in more detail. He put in $108 over the year and then took out
$120, so he earned $12 ‘interest’ ($120 minus 108). He had on average $54 ‘on deposit’ during that
year
13
. So he earned $12 on $54, that is a rate of just over 22% a year. Not bad.
The first member also put in $108 over the year, but, as we saw, she took out $96 on the very first
day. So she ‘paid’ $12 in ‘interest’ (matching what the last member ‘earned’). Since she paid in an
average of $9 a month she had ‘repaid’ her $96 ‘loan’ in a little under eleven months. She thus paid
$12 interest for a loan that averaged $48 over eleven months. This is an interest rate of 24% over
eleven months, or about 26% a year.
Here are the diagrams for the first and last members in our example, at the same scale:

12
In some auction ROSCAs only the nine non-winners would share this discount.
13
Well, actually, not quite, since he put in more in the last four months than he did in the first four. This will skew
things slightly in his favour compared to my calculation above. The inverse is true for the first member.
xxv
My suggestion that the
auction is a way of ensuring
that the lump sums go,

each round, to those who
most need them sometimes
provokes strong
disagreement. ‘Not so’, say
these critics, ‘as in much of
the real world, the sums go
not to those who most need
them but merely to those
who can most easily afford
them. In this way they
merely perpetuate the
conditions that the poor
unfairly suffer in so many other aspects of life’. But whatever may be the truth of that as a general
commentary on life, it isn’t really true in the case of an auction ROSCA. After all, even the poorest
member of all can still bid in the first round, and can win it if he is willing to accept the biggest
discount. He isn’t disadvantaged by a richer member standing next door to him with his pockets
bulging with cash.
The range of bid sizes in auction ROSCAs
The rates in the example above - 26% a year for a loan and 22% a year on savings - would, in most
countries, be more attractive than most other services available to poor savers and borrowers. But
these rates are not typical for ROSCAs, they are merely examples to demonstrate the arithmetic
involved. In practice ROSCA members often bid much more than the modest 20% of the lump sum
on offer used in the example above. In northern coastal Vietnam I talked to capital-hungry fishermen
eager to invest in new equipment and found that in their ROSCAs, which are very common, early
bids commonly reach 50% of the lump sum, or more
14
. In other countries, notably India, bids are
often so high that government has tried to legislate to limit them. Very high bidding means that net
‘borrowers’ pay a higher price for their ‘loans’ while those who choose to take their pay-out near the
end receive very high implied rates of interest on their savings. ROSCAs are thus a very sensitive

instrument for measuring, at frequent intervals, the price to the poor of capital in a local area (a
point that economists and the designers of financial services for the poor might note).
The ROSCA ‘sprint’
Sukhwinder Arora noted that in the Indian towns that we were studying many slum dwellers were
pushing money through ROSCAs (particularly auction ones) at a much faster rate than through any
other type of savings club or financial service. He rightly describes ROSCAs as ‘sprints’, comparing
them to more sedate services such as a savings bank, which he calls ‘marathons’. In an ordinary
savings account at a bank or Post Office you build up your savings gradually, over the long term,
and it doesn't matter much if you don't save for several weeks or even months on end. In an auction
ROSCA, by contrast, you commit yourself to the highest possible level and frequency of regular
saving you dare, by joining the ROSCA with the biggest contributions and most frequent rounds you
can find (or that will let you in). For that reason the very poorest are the least well represented among
users of auction ROSCAs. We noted in the first chapter that one disadvantage of devices which
require fixed contributions at fixed intervals is that the very poor may be scared off or prevented from
joining, because of fear of not being able to maintain the strict schedule.
As you would expect, people with businesses favour auction ROSCAs as a way of getting hold of
investment capital. Equally, people with regular incomes - above all salaries - favour them as a way
of getting a good return on their regular savings. Businessmen can be fairly sure of being able to
make the contributions at the fast pace required, and their businesses represent for them attractive
opportunities for investment of the lump sum. Because in many societies running a business is seen
as a male activity, auction ROSCAs are sometimes seen as ‘men’s ROSCAs’, while lottery ROSCAs are

14
Of course, members who bid high also, like the net savers, enjoy high returns on their deposits, thus offsetting
their costs somewhat.
CHART SEVEN:
AUCTION ROSCA
lump sum lump sum
pay-ins: same for both